A structured-similarity knowledge graph (Tier 1). Edges are mutual nearest-neighbours — two companies link only if each is in the other's top matches — so genuinely-distinct companies float free rather than being forced to connect. Colour = detected community; bigger nodes are hubs. Drag to move, scroll to zoom, click a node to frame its cluster; click empty space to see the whole map. Hover a node for its product, or a cluster label for its digital×chain make-up, displacement-mode mix, and closest clusters.
How each cluster splits across the four displacement verbs — the way its companies act on the incumbent value chain. Circle area ∝ number of companies (scaled to the busiest cell); colour marks the verb, columns run augments→replaces.